Journal of Archaeological Science (1996) 23, 689704

Oxygen Isotopic Analysis of Archaeological Shells to Detect

Seasonal Use of Wetlands on the Southern Pacific Coast of
Mexico
Douglas J. Kennett
Department of Anthropology, University of California, Santa Barbara, CA 93106, U.S.A.

Barbara Voorhies
Department of Anthropology, University of Colorado, Boulder, CO 80309, U.S.A.
(Received 31 December 1994, revised manuscript accepted 21 August 1995)
Oxygen isotopic ratios in modern and archaeological marsh clam shells (Polymesoda radiata) from the Acapetahua
Estuary in southwestern Mexico record large scale salinity fluctuations caused by alternating wet and dry seasons. Thus,
prehistoric patterns of rainfall can be reconstructed and the season of molluscan death can be estimated. Changes in the
oxygen isotopic patterns preserved in marsh clam shells from late Archaic period (c. 30001800 ) archaeological
deposits indicate that the season of shellfish harvesting changed dynamically through time in this region. Based on this
study, and other lines of archaeological evidence, we argue that huntergatherers during the early stages of the late
Archaic period visited locations in the littoral zone throughout the year with a focus during dry season months.
Through the late Archaic period a general trend occurred toward wet season use of these locations. This culminated at
the end of the late Archaic period with the exclusive use of the littoral zone during wet season months. These data
indicate a fundamental shift in the way these estuarine locations were being used. We argue that people living in this
region altered their overall subsistence strategy during the late Archaic period due to scheduling conflicts that occurred
with the adoption of maize agriculture.
? 1996 Academic Press Limited
Keywords: MESOAMERICA, ARCHAIC PERIOD, COASTAL HUNTERGATHERERS, ORIGINS OF
AGRICULTURE, SHELL MIDDEN, ISOTOPE ANALYSIS.

prehistoric people living in the region during this time

interval. These data are assessed with other lines of
archaeological evidence to better understand the season in which these sites were occupied and how they
functioned within a larger subsistence and settlement
system.
The late Archaic period in Mesoamerican prehistory
is of special significance because it is the time when
ancient people were making an economic transition
from foraging to farming. It is also the period immediately preceding the development of greater sociopolitical complexity. Understanding the role of coastal
resources to the prehistoric huntergatherers in the
Mesoamerican lowlands prior and during the introduction of agriculture may assist with the understanding of
this important transformation. Our principal finding is
that changes occurred in the use of estuarine resources
throughout the late Archaic period. These changes
were probably associated with scheduling conflicts that
developed with the intensification of maize agriculture
in the region.

Introduction
econstructing changes in human subsistence
and settlement patterns is an integral part of
understanding how prehistoric people adapted
to their social and natural environments through
time. One component of reconstructing prehistoric
subsistence and settlement is accurately determining
the seasonality of resource and site use. Evidence for
the seasonal use of resources and sites range from the
simple presence or absence of seasonally available
plants and migratory animal species (Monks, 1981;
Deith, 1985) to more complex studies of bone and
antler or tooth and shell growth increments (Deith,
1985; Bernstein, 1990; Lieberman, 1993).
The purpose of this paper is to discuss the seasonal
use of littoral sites by huntergatherers who lived along
the Pacific coast of southern Mexico during the late
Archaic period (30001800 ). Oxygen isotopic analysis of estuarine mollusc shell carbonate is used to
reconstruct the season when clams were harvested by

689
0305-4403/96/050689+16 $18.00/0

? 1996 Academic Press Limited

690 D. J. Kennett and B. Voorhies

Study Area
The environment
The late Archaic period shell deposits discussed in this
paper are located on the Pacific coastal plain of southern Mexico (Figure 1). This region was known as the
Soconusco to the Aztecs, who traded with and exacted
tribute from its people, and the term continues to
be used today to refer to the strip of coast between
the town of Pijijiapan and the MexicanGuatamalan
border (Voorhies, 1989).
Rivers originating in the Sierra Madre de Chiapas,
the mountain range flanking the coast, transect the
coastal plain and flow into the coastal lagoons,
swamps, and estuaries of this region. The canals and
shallow water lagoons of the Acapetahua Estuary are
surrounded by a patchwork of mangrove forest and
cattail marsh. This productive aquatic habitat supports
a wide range of marine and estuarine organisms including fish, shrimp, molluscs and waterfowl. Faunal and
floral remains found in the late Archaic period shell
mounds in the estuary suggest that the environment
was basically similar to that of today (Voorhies, 1976;
Hudson, Walker & Voorhies, 1989). Phytolith studies
indicate that before 5000 years ago the coastal plain
was covered by a tropical deciduous forest that was
subsequently disturbed (Jones, 1988; Jones & Mora,
1993).
Southwestern Mexico is influenced by highly seasonal tropical monsoonal rains (Viv Escoto, 1964).
Rain falling on the Soconusco coastal plain and the
Sierra Madre de Chiapas provides a seasonally variable
supply of fresh river water to the brackish water of the
estuarine system. Large volumes of fresh water flow
into the littoral zone during the wet season between
April and October, with a much smaller influx during
dry season months between November and March.
Stable carbon and oxygen isotopes indicate that this
tropical monsoonal rainfall regime persisted throughout the late Archaic period (Kennett & Voorhies,
1995).
The Chantuto people
Late Archaic period occupation of the Pacific coastal
plain coincides with the stabilization of global sea level,
approximately at 6000 years (Fairbanks, 1989).
Human populations inhabiting the Soconusco region
at this time have been named the Chantuto people
(Voorhies, 1976); evidence for their existence consists
of six large shell mound sites. Five of these sites are
located in the Acapetahua Estuary (Drucker, 1948;
Lorenzo, 1955; Voorhies, 1976; Navarrete, n.d.). The
other site, Cerro de las Conchas, is situated near the
inland margin of the El Hueyate swamp adjacent to
the estuary to the southeast (Clark, 1994) (Figure 1).
Although these sites represent the earliest recognizable
human occupation on the coastal plain, earlier deposits
were probably covered by the Late Pleistocene/Early

Holocene marine transgression or have been obscured

by sediments associated with this actively prograding
coastline.
Cerro de las Conchas, the oldest of the six late
Archaic period middens in the Soconusco region, was
deposited between 4000 and 3000 (Blake et al.,
1995; Clark, 1994). This shell midden is approximately
80 m in diameter and over 5 m high. Although this
deposit is presently located near the freshwater El
Hueyate swamp, faunal remains at the site indicate
that it was once adjacent to the littoral zone. The
deposits are dominated by marsh clams; however, a
lense of limpets also was discovered at the site (Clark,
1994). Although this site is considered to be an early
manifestation of the Chantuto phase cultural complex,
it is not the focus of this study.
This study focuses on the five shell mounds located
in the Acapetahua Estuary (Figure 1) that were formed
during the millennium after Cerro de los Conchas, that
is, during the third millennium before the current era
(Blake et al., 1991). All of the shell mound sites in the
Acapetahua Estuary are remarkably similar (Voorhies,
1976), consisting almost exclusively of the marsh clam
Polymesoda radiata (Severyn, 1993),* lying in alternating beds of crushed and whole shells. These aceramic
shell deposits, dating to the late Archaic period, are
overlain by dark brown soil deposits with diagnostic
ceramics from later time periods.
Clamshells used for this investigation are from the
site of Tlacuachero, excavated by Voorhies in the early
1970s (Voorhies, 1976) and again in the late 1980s
(Michaels & Voorhies, 1989). Tlacuachero forms an
artificial island within the wetlands that is approximately 140#128 m in size and greater than 74 m in
height (Voorhies, 1976) (Figure 2). The predominant constituents of the lower aceramic deposits at
Tlacuachero are shells of P. radiata (Figure 3), comprising 9955% of the faunal assemblage (Voorhies,
Michaels & Riser, 1991). Other invertebrate and vertebrate remains that comprise the faunal assemblage are
from species common in the present day estuarine
environment (Voorhies, 1976; Hudson et al., 1989;
Anikouchine, 1990). Of these, fish are the dominant
vertebrate species represented (Anikouchine, 1990).
Open water marine vertebrate and invertebrate species
are notably scarce in the Chantuto phase layers.
The stratigraphy of the lower, aceramic component
at Tlacuachero, and the other late Archaic period shell
mounds, is characterized by distinct alternating beds of
burned and unburned shell (Figure 4). Laterally extensive layers of unburned, whole P. radiata shells are
interbedded with equally extensive layers of burned,
broken shells of the same species. Lenses of dumped
*This clam was formerly called Neocyrena ordinaria but its taxonomic nomenclature recently has been revised.
We returned to the site more recently, in 1994, during which time
additional excavations were made and samples of clam shell and
charcoal collected for further analysis. However, these samples were
not used in the study reported here.

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 691

bo
s
Bo

Study area

R.

ov
i

lle
ro

.S
es
ec
ap
a

R.

R. M

arg
ari t
as

apa
R. C
o

R.

Pi
jij
ia
pa
n

Pijijiapan

Vuelta
Limon
ta
ia
n
lu
ar
ca
Sa
.
a
~na M
R
C
.
R
Do
L. Chantuto
R.
s

es

po
bl
ad
o

paluta
R. Za

R.
D

R. a
ap
Fil
o

nch

oA

d
Va
R.

R.
Comaltitlan

Cerro de las
Conchas

L. Panzacola

R. Huixtla

Islona
L. el Campon
Chantuto
Campon
L.
Tlacuachero
Teculapa
El Chorro
Zapotillo

R. Cin

Pacific Ocean

talapa

la

co

Ni

pa

ma

R.

Isla

ue

lH

E
etan

te

ya

R. Huehu

R. P

ump

uapa

N
5 3 10

km

10

20

R.

ata

Co

Figure 1. Study area, showing late Archaic period shell mound sites, coastal lagoons of the Acapetahua Estuary, the El Hueyate swamp and
coastal rivers discussed in this paper. : Freshwater swamp; /: Archaic period shell mounds; 4: Archaic period-site.

material (basket dumps), often found in trash midden

contexts, are rare, as are pits, postmolds and other
features. We believe that these bedded deposits are the
actual remains of clam bakes, that appear to have been
carried out by placing a large bed of clams over a bed
of coals.* These blanket type middens (Waselkov,
1987) are also marked by a low diversity of stone
tools. Only a small number of millingstone artefacts,
hammerstones, and obsidian flakes have been found in
the lower deposits at Tlacuachero.
It is likely that implements used for exploiting estuarine resources were made of perishable materials that
have not survived archaeologically. The diameters
of fish vertebrae at Tlacuachero indicate that the
Chantuto people procured many small fish (Voorhies
et al., 1991), that were less than 15 cm long. This
suggests to us that late Archaic period people used fine
mesh baskets or nets to catch fish and probably
shrimp. Similarly, the small size of some marsh clams
*Voorhies, Michaels & Riser (1991: 2324) had earlier interpretated
the bedding of the shell deposits as having been caused by intentional
burning of vegetation when the site was reoccupied after a period of
abandonment. While this hypothesis has not been disproven, we now
favour the view that the bedding at the site occurred as a result of
cooking the clams.

harvested (<1/4+) indicates that a mass harvesting

device, such as shovel or rake with closely spaced tines,
was used to collect these bivalves en masse (Voorhies,
1976).
Information about sites on the Soconusco coastal
plain contemporary with the shell mounds in the
littoral zone is scant. One site, Vuelta Limn (Figure
1), is located upstream from the Acapetahua Estuary,
where it was discovered by us in a river bank (Voorhies
& Kennett, 1994). There we excavated a trash midden
containing many fire cracked and waterworn rocks, as
well as some stone tools. We think that this site
provides evidence of a sedentary (or semi-sedentary)
occupation, which Voorhies expected on the basis of
her theoretical model of coastal adaptation during the
late Archaic period (Michaels & Voorhies, 1989;
Voorhies, 1991).

Seasonal Mobility on the Soconusco Coastal

Plain: the Model
Various settlement and subsistence strategies are employed by huntergatherers faced with the challenges
of exploiting seasonally available resources. Humans

692 D. J. Kennett and B. Voorhies

0
N

1
2
3
4
5

24B
18B

AB13
CD0

12B

AB10

6B
N0B2

6
42C

34C

38C 24C 18C 12C 8C

6D 12D 18D 24D

38D

34D

3 2 1

6A
12A
15A

IJ4
KL6

S16W1

5 CH0
4
3

IJ9

EF9
GH6

4 5

2
1
0

m
0

20

40

Figure 2. Plan of excavation areas and auger test locations of Tlacuachero (site CAP-7), a shell mound located in the Acapetahua Estuary.
Excavation units and auger tests from the 1973 and 1988 field seasons are shown (adapted from Michaels & Voorhies, 1989). : Limit
of clay floor; /: 1973 excavations; .: 1988 excavations; -: auger locations. Contour interval=05 m above sea level; declination: 8 degrees
East.

respond to changes in the abundance and distribution

of natural resources and to other non-economic factors
(Hard & Merrill, 1992). Archaeologists are faced with
trying to reconstruct these dynamic patterns of prehistoric movement with the static archaeological
record (Binford, 1983). To this end, Binfords (1980)
foragercollector continuum has been useful for building heuristic models of prehistoric subsistence and
settlement that can be tested against the archaeological
record. This continuum, based on extensive ethnographic research, ranges from highly mobile foragers to
relatively sedentary logistical collectors.
The forager strategy is generally practiced by
huntergatherers living in highly productive, homogeneous environments where resources are abundant and
continuously available throughout the year. Generally,
these small, highly mobile, groups move from habitat
to habitat throughout the year in synchrony with the

Figure 3. Two valves of a modern marsh clam, Polymesoda radiata

(Photograph by B. Voorhies; scale in cm).

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 693

Figure 4. Photograph of bedded stratigraphy at Tlacuachero (CAP7). Notice layers of unburned whole clamshell valves interbedded
with layers of burned broken shells (Photograph by B. Voorhies).

seasonal distribution of the resources. During this

seasonal migration, foragers establish a residential base
in a resource zone to gain access to resources within a
2 h walk of the encampment. Once the resources in this
zone are depleted, the group moves to a different
location. Binford (1980) argues that because there are
no seasonal food shortages there is no reason to store
resources. Most food is prepared for immediate
consumption, and processing occurs at centralized
residential base camps.
Two settlement types are generated by this type of
subsistence-settlement strategy: residential bases and
locations. The size of residential bases depends on the
number of occupants and the intensity and length of
occupation, but, in general, these sites are more ephemeral than the base camps of collectors. Highly mobile
groups deposit relatively homogenous refuse that is
functionally related to the activity performed in each
habitat (Bettinger, 1991). Most of the processing,
manufacturing, and tool maintenance occurs at the
base camps because they are the loci of all subsistence activities.Task groups are deployed to locations
to exploit plants and animals. These locations generally show evidence of a single extractive task (e.g.
butchering).
Conversely, a collector strategy is employed by
huntergatherers in environments where the distribution of resources is spatially and temporally variable. Collectors inhabit semi-permanent encampments

and extract resources with planned logistical forays to

different environmental zones. Seasonal shortfalls in
resources that cannot be easily mitigated by periodic
movement require advanced planning and more complex strategies for survival. Collectors intensely exploit
seasonally available resources and store the excess food
yields for times when productivity declines. Storage
extends the temporal availability of resources but reduces a groups mobility so that shortfalls that were
once solved with periodic movement become more
costly. A greater degree of internal specialization and
coordination is needed to develop extraction techniques to procure seasonally available resources in
bulk. Special task groups are deployed to collect specific resources for a group as a whole. These activities
usually involve the development, curation and maintenance of specialized tools that make extracting resources more efficient (Torrence, 1983; Bamforth,
1986). Sometimes large, bulky equipment is cached
close to the location of the resource. Base camps are
usually centrally located, and task groups travel from
this locus to hunt or collect in different resource zones
and return to the base camp.
Archaeologically, the assemblages found at the residential bases of collectors are more heterogeneous than
forager base camps, reflecting a greater diversity in
activities carried out over a larger area. Evidence for
storage is also a good archaeological indicator of this
more logistically organized subsistence strategy. In
addition residential bases and locations, collectors also
leave a greater variety of site types including field
camps, caches and stations for gathering environmental
information. Locations used by collectors generally
exhibit evidence for the procurement and processing of
resources in bulk.
It is probable that people living in the coastal
lowlands of Mesoamerica during the late Archaic
period were primarily hunters and gatherers, although
we lack a comprehensive picture of their subsistence
economies. Michaels & Voorhies (1989) postulate that
in addition to hunting and gathering wild food, coastal
dwellers in the Soconusco were practicing some form
of early horticulture. Regional settlement data of the
type necessary to establish how these huntergatherers
were moving around the landscape are limited. Based
on the available evidence of coastal populations
in Mesoamerica, Michaels & Voorhies (1989) have
developed a heuristic subsistence-settlement model for
human populations who lived in the Mesoamerican
lowlands during the late Archaic period. In this model,
the organization of subsistence and settlement strategies resembles collectors rather than foragers. This
hypothetical settlement system consists of one or two
relatively sedentary base camps that were used as loci
for logistically organized groups that exploited resources from different habitats on the coastal plain and
adjacent mountain ranges.
Large residential bases were probably located on the
coastal plain (Michaels & Voorhies, 1989), and we

694 D. J. Kennett and B. Voorhies

think that Vuelta Limn is the first such site to be

found. The Chantuto phase shell mounds are interpreted as specialized locations for extracting shellfish,
shrimp, fish and other estuarine resources. This interpretation is supported by (1) the great areal extent
and depth of the deposits, indicating intense exploitation of estuarine resources beyond the need of the
group collecting them, (2) the distinct bedded stratigraphy of shell mound sites, suggesting periodic visits
to these locations, (3) the undisturbed nature of the
deposits, indicating short term occupation, (4) the
virtual absence of evidence for structures* and domestic cooking areas including postmolds, floors, pits or
hearths, and (5) the limited range of unspecialized
stone tools.
A survey of the present day seasonal availability of
resources in the Acapetahua Estuary suggests that the
dry season is the optimal time for people to exploit
wild resources from the littoral zone. Although most
plants and animals in this aquatic habitat are available
throughout the year, there is a resource pulse starting
in the early dry season and continuing to the beginning
of the wet season. Migratory waterfowl appear in the
region during the late wet season/early dry season. This
is followed by a peak in shrimp availability during the
dry season, of great importance to present day populations living in the estuary (Voorhies et al., 1991).
Female turtles, crocodiles, and iguanas also carry
their eggs at this time. Finally, there may be a general
reduction in terrestrial biomass during the dry
season that would have made the littoral zone of the
Acapetahua estuary especially attractive during that
time of the year.
Based on the available archaeological evidence
it appears that the Chantuto phase shell mound
sites were occupied periodically, possibly seasonally,
throughout the late Archaic period by logistically
organized collectors. Determining precisely when the
littoral zone was occupied during the annual cycle is
one component in deciphering this ancient pattern of
settlement. It may also provide insight into why agriculture was adopted on the coastal plain and why the
hunting and gathering subsistence strategy of the late
Archaic period cultural complex came to an end.

Seasonal Determinations with Oxygen

Isotopic Analysis: the Test
Oxygen isotopic analysis of molluscan shell carbonate
is a well established technique for determining the
season of prehistoric shellfish harvesting. The method
was initially recognized as a powerful tool for palaeoenvironmental reconstruction because oxygen isotopic
ratios in calcareous fossils contain information about
the physical and chemical environment of their growth
(Wefer & Berger, 1991). Two environmental factors
*Evidence for structures is present on the one clay surface that was
discovered at Tlacuachero but this floor construction and the structures that were upon it, appear to be unique (Voorhies et al., 1991).

contribute to the isotopic composition of shell carbonate: the isotopic composition of seawater and water
temperature. Urey (1947) showed that the stable
oxygen isotopic composition of calcium carbonate
deposited by marine molluscs was temperature dependent and thus of great value as a palaeothermometer.
A paleotemperature equation was developed by
Epstein et al. (1951, 1953) based on oxygen isotopic
measurements of mollusc shell carbonate precipitated
at known water temperatures. In the equation:
T=AB(cw)+C(cw)2
T is equal to temperature in )C and A, B, C are
constants respectively equalling 164, 42 and 013. The
symbol c is the oxygen isotopic ratio of the carbonate,
expressed as a deviation in (ppm) from a standard
carbonate.
c =

18O sample
"1 # 1000
18O standard

w represents the oxygen isotopic composition of the

water expressed in a similar fashion, as a deviation
from standard mean ocean water (smow). In order to
solve the equation for temperature (T), the delta value
of the water must be known. When the delta value for
the water (w) is constant, the oxygen isotopic ratio
increases by approximately 02 for every 1)C increase
in water temperature.
Thus, in the open ocean, where the composition of
sea-water has been relatively stable since the middle
Holocene (26000 years to present day), oxygen isotopic measurements of calcium carbonate extracted
from the sequential growth increments of molluscs
reflect seasonal fluctuations in water temperature and
the season of death can be estimated. Shackleton
(1969) was the first to point out the applicability of
this technique for archaeologists interested in determining the season of mollusc collection from shells in
archaeological deposits.
The methods for extracting seasonal information
from mollusc shells were initially worked out by
Shackleton (1973) and have changed little since. Calcium carbonate samples are extracted along a shells
growth axis from the growth margin towards the hinge.
Using specimens of Patella tabularis collected alive
from Nelsons Bay Cove, South Africa, Shackleton
determined that oxygen isotopic changes through the
growth of mollusc shells paralleled seasonal fluctuations in temperature and that the shell margin samples
accurately reflected the season of molluscan death.
Shells from prehistoric midden deposits, dating between 9000 and 5000 years ago, in the region indicated
that molluscs were harvested primarily during winter
(cold weather) months.
Based on this study, Shackleton (1973) outlined a
number of criteria that should be met to make seasonal
temperature determinations using oxygen isotopic
analysis. First, shell growth must take place under

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 695

conditions of isotopic equilibrium with the surrounding water. Second, the isotopic composition of the
water in which the shellfish lives must remain constant
throughout the year. Third, the shell must precipitate
carbonate throughout the year at a relatively fast rate.
Finally, the seasonal temperature range must be greater
than the week-to-week variations in temperature.
Post depositional diagenesis of shell carbonate
presents a potential problem to archaeologists applying
this technique to shells from prehistoric midden deposits. Shell surface carbonate is particularly susceptible to dissolution and recrystallization (Shackleton,
1973; Bailey, Deith & Shackleton, 1983), and chemical
exchange with percolating ground water can alter the
shells original isotopic signature. In some depositional
contents diagenesis is expedited by endolithic bluealgae that bore into the surface of the shells (Deith,
1985). Oxygen isotopic values in these cases will be
lowered, because ground water has lower 18O values
(i.e. relatively more 16O) than ocean water. Establishing the isotopic systematics if living molluscs of the
same species provides one baseline for evaluating the
integrity of the isotopic composition of archaeological
samples.
More recent literature has focused on establishing
the precision of the oxygen isotopic method for determining seasonality. Based on a study of modern and
archaeological Myrtilus californianus specimens from
the California coast (Killingley, 1980, 1981; see also
Killingley & Berger (1979) and Glassow et al. (1994))
proposed that the month of prehistoric shellfish collection can be determined by statistical treatment of
oxygen isotopic data. Bailey et al. (1983; also Deith
(1985)) argued, in contrast, that determining the season
of molluscan death to the month was unrealistic because of known oxygen isotopic differences between
species and regional climatic variation through time.
Oxygen isotopic analyses of marine molluscs have
been successful because the isotopic composition of
ocean water has remained relatively constant since the
middle Holocene, whereas variations in temperature
cause predictable changes in isotopic composition.
However, this is not the case in coastal estuaries where
large fluctuations occur in isotopic composition of
estuarine waters. This dynamic situation is largely
caused by changes in the influx of river water, as well as
by seasonal temperature changes. The influx of river
water causes significant changes in 18O composition
of estuarine waters because fresh water from the continent has relatively low 18O values (Keith, Anderson
& Eichler, 1963). This complex interaction may cause
difficulties in deciphering the environmental causes
of oxygen isotopic change in estuarine molluscs,
especially in temperate regions where there are large
seasonal changes in temperature. Deith (1983, 1988)
argued that oxygen isotopic analysis is best suited for
open ocean marine molluscs, but some estuarine forms
are suitable if they retain a temperature-dependent
signal.

Little oxygen isotopic work has been carried out

using tropical molluscan species, because large seasonal differences in temperature do not occur. Tropical
estuaries, however, provide a context in which water
temperature remains relatively constant throughout
the year, while seasonal fluctuations in freshwater
runoff cause distinct changes in the oxygen isotopic
composition of the water, overwhelming the temperature effect (Kennett & Voorhies, 1995). In these settings, oxygen isotope analysis of molluscan shells is
potentially a powerful tool for reconstructing ancient
water regimes and determining the season of shellfish
harvesting.
The rainfall regime along the southern Pacific coast
of Chiapas is highly seasonal. The Soconusco region
receives an average rainfall of 3200 mm (based on
measurements taken by the Comisin Nacional de
Agua at Escuintla between 1975 and 1990). Much of
this rain falls between April and October, with negligible amounts falling during the rest of the year.
During the rainy season, the rivers that transect the
Soconusco coastal plain flood the coastal estuaries with
fresh water. Kennett & Voorhies (1995) hypothesized
that during wet season months, waters of the
Acapetahua Estuary would exhibit lower 18O values
compared with the dry season, and that this seasonal
variability should be recorded in the shells of the marsh
clam Polymesoda radiata. If such relations existed
throughout the Holocene, they should provide a basis
for determining the seasons of collection by prehistoric
populations living in the Soconusco region during the
late Archaic period. This will, in turn, provide valuable
information about site function.

Materials and Methods

Modern clam shells
Living specimens of P. radiata were collected from the
Acapetahua Estuary to better understand the relationship between fluctuating environmental factors and the
oxygen isotopic variation in the shells. Monthly collections of these clams were taken from the Los Cerritos
Lagoon (see Figure 1) during the course of one annual
cycle spanning the calendar years 1989 and 1990,
excluding the months of June and July. Water samples
were collected from the lagoon at the same time.
Five right valves were randomly selected for analysis
from each of the 10 monthly collections. The shells
were thoroughly cleaned to remove any extraneous
organic material that might contaminate the sample.
Any obvious organic material on the surface of the
shell, including the periostracum, was removed with a
razor blade. The shells were then rinsed in deionized
water and baked in an oven at 85)C until they were dry
(Killingley & Bergery, 1979). It was often necessary to
repeat this process until all of the visible organic
material was removed.

696 D. J. Kennett and B. Voorhies

1m

Dark brown soil

w/ ceramic inclusions

1.201.40 m

Early Preclassic-Postclassic

Period

Levels sampled for

stable isotope study

2.202.40 m

Calibrated radiocarbon
date-range (cal. BC)

224 BCAD 3

27572464
25642342

4.404.60 m

Clay floor

Late Archaic

3.303.40 m

34983145

30942902

5.405.60 m

33382926
6.006.20 m
30372788
29132627
7.007.20 m
Legend
Whole shell
Burned crushed
shell and carbon
Figure 5. Idealized stratigraphic profile of excavation unit NOE2 at Tlacuachero (CAP-7).

Archaeological clam shells

The archaeological specimens of P. radiata analysed
are from the site of Tlacuachero (CAP-7). In the field,
10 cm3 of matrix were collected at intervals of 10 cm
(corresponding to the lower 10 cm of each 20 cm
excavation level). The clam valves used in this study
were removed from these matrix samples, but only
shells from seven levels were analysed, the levels being
separated by approximately 1 m intervals (Figure 5).
The highest group of samples came from the top of the
late Archaic period deposits 120140 m below the
surface whereas the lowest group of samples came from
the 700720 m level below the surface.

Several methods were used to establish that there has

been no postdepositional alteration of the archaeological shell carbonate. Thin sections of modern and
archaeological shells were first examined with conventional and scanning electron microscopes. Visual
inspection of the crystalline structure of the archaeological shells indicated that they are well preserved.
Modern and archaeological shells also were examined
using X-ray diffraction. The crystal lattice of contemporary P. radiata shells consists of aragonite. We
determined that the primary aragonite in the archaeological specimens is intact and that secondary deposits
of calcite, indicative of diagenesis, are absent.

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 697

Figure 6. Drawing of Polymesoda radiata shell showing how successive growth increments were sampled for oxygen isotopic analysis (Drawing
by Meredith Kennett).

10

9
DB
m 18O
ar
gi
n
(P

)
ow
(sm

er

Sh

ell

W
at

18O isotope (ml1)

4
3
2

S O N D J F M A M J J
Months of the year (19891990)

Figure 7. Oxygen isotopic composition of modern shell margins

(N=5 per month) plotted against the 18OSMOW of estuarine water
samples collected at the same time from the surface of the Los
Cerritos Lagoon.

Twenty archaeological clams were selected from

each level sampled for oxygen isotopic analysis. Each
valve was rinsed with deionized water to remove the
adhering soil matrix and then dried in an oven at 85)C.
The surface of each archaeological shell was etched
with a weak solution of HCl to remove any possible
diagenically altered carbonate (Bailey et al., 1983).
After the outer fraction of the shell was removed, the
samples were rinsed again with deionized water and
dried at 85)C.
The margin of each modern and archaeological shell,
representing the final stages of growth prior to harvest-

ing, was removed with a sterile razor blade. Several

modern (N=6) and archaeological (N=14) shells were
further sampled (Figure 6) to reconstruct the oxygen
isotopic variation throughout the life of the molluscs.
Samples were extracted in 15 mm increments along the
growth axis of the shell using a 05 mm dental drill.
Each sample was ground into a fine powder with a
mortar and pestle, loaded into a small copper vessel
and roasted, under vacuum, at 350)C for one hour.
The roasting process oxidizes any remaining organic
material. After roasting, the pure calcium carbonate
sample was reacted in orthophosphoric acid at 90)C.
The oxygen isotopic ratios of the resulting CO2 were
determined by mass spectrometry (Finnegan MAT-251
Mass Spectrometer). All measurements are expressed
as a deviation from an internationally accepted standard, PeeDee belemite, a carbonate fossil from South
Carolina (Herz, 1990). The precision of the oxygenisotopic ratios is 01%. More negative values indicate
higher proportions of the lighter 16O isotope compared
to the heavier 18O isotope and vice versa.

Results
The range of 18O variation of the monthly modern
shell margin samples from September, 1989 to August,
1990 is shown in Figure 7. This represents an annual
18O range of 4 (2"8 to "9 between the
months of July and January and 2"5 to "6
between the months of February and June). The
annual range of oxygen isotopic composition of shell
margin carbonate is greater (4) than that of any
single month (1), thus meeting Shackletons criterion. Moreover, comparing both lines in Figure 7 it
can be seen that the 18OPDB of the shell margin

18

m 18
ar O
gi
n
(P
Sh
el
l

3 1

B)

O isotope (ml )

18

0
S O N D J F M A M J J
Months of the year (19891990)

10

inf

Sh
(P ell
D m
B) ar
g

all

in

500

300
200

18

400

Rainfall (mm)

(mm

Ra

18O isotope (ml1)

600

10

700

20

Figure 10. Oxygen isotopic composition of margins of modern

shell plotted against water flow in the Cintalapa River (Comisin
Nactional de Agua) between September, 1989 and August, 1990.

800

10

5
4

Figure 8. Oxygen isotopic composition of modern shell margins

plotted against the salinity of estuarine water samples taken from the
surface of Los Cerritos Lagoon.

30

S O N D J F M A M J J
Months of the year (19891990)

40

lape

inta

50

w (C

18

Salinity

18O
Shell margin

60

O isotope (ml )

Salinity ()

10

r flo

24
22
20
18
16
14
12
10
8
6
4
2
0

Rive

O isotope (ml )

10

Water flow (m s )

698 D. J. Kennett and B. Voorhies

8
6
4

0
A

Figure 9. Oxygen isotopic composition of margins of modern

shell, plotted against average monthly rainfall measured on the
coastal plain at Escuintla (Comisin Nacional de Agua), between
September, 1989 and August, 1990.

carbonate parallels the 18OSMOW of the water samples

collected each month from the Los Cerritos Lagoon.
Oxygen isotopic composition of the estuarine water is
higher between January and June (2"3 to "5%) and
lower from July to December (2"61 to "78).
Successive changes in the oxygen isotopic composition of both shells and water are inversely correlated
with salinity fluctuations in the Acapetahua Estuary
(Figure 8). The oxygen isotopic values of the estuarine
water (Figure 7) and the shell margin carbonate (Figures 7 and 8) are higher (2"42 to "62) when the
water is more saline (1820) between February and
June (dry season). Oxygen isotopic values are consistently lower (2"7 to "92) between July and
January when the estuarine waters are almost fresh
(12).
Figure 9 compares the oxygen isotopic composition
of shell margin carbonate and rainfall for the same
interval of time. The total rainfall recorded at the
Escuintla Meteorological Station on the coastal plain
between September 1989 and August 1990 was
3370 mm, and ranged from 700 mm in September 1989
to 48 mm in January of 1990. Although similar trends

is

Polymesoda radiata
Modern specimen # 94
Collected-August 8, 1990

2
0

18 O

(P

ot

Shell margin

100
S O N D J F M A M J J
Months of the year (19891990)

B)

pe

9
6
12
15
18
21
Distance from shell margin (mm)

24

Figure 11. Oxygen and carbon isotopic profiles of a modern clam

shell collected from the Los Cerritos Lagoon in August, 1990.

are exhibited between changes in 18O values recorded

in shell margins and rainfall, a lag of 23 months is
apparent between the two signals (Figure 9). This lag is
also apparent between changes in the oxygen isotopic
composition of shell carbonate and the velocity of the
Cintalapa River flowing into the Los Cerritos Lagoon
(Figure 10). During this particular annual cycle (1989
1990), water velocity was highest (55 m3 s "1) during
September, 1989. Water flow decreased rapidly after
September and remained low (24 cm3 s "1) until
March of 1990, when it began to increase again.
Oxygen isotopic variation in the growth increments
of all shells analysed record successive fluctuations in
estuarine salinity. The amplitude of oxygen isotopic
change in modern shell profiles is consistent with the
range exhibited in the shell margin samples from
molluscs collected at month intervals. An oxygen isotopic profile for a modern marsh clam collected in
August of 1990 is displayed in Figure 11. The amplitude of oxygen isotopic change in the two years of
growth represented by this profile is 6, ranging from
"35 to "95. The oxygen isotopic value at the
shell growth margin is "70, which is consistent
with values (see Figure 7) from other specimens of P.
radiata collected from the Los Cerritos lagoon during
August of 1990.

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 699

Wet

11

18

Level 1.201.40

top
e(
PD
B)
iso

9
18
O

Modern

Polymesoda radiata
CS-7/NOE 2
Level 1.201.40 m
Sample # 140

Shell margin

O isotope (ml )

Dry

7
5
3

Level 2.02.40

3
6
9
12
15
Distance from shell margin (mm)

Figure 13. Oxygen isotopic profile of an archaeological clam shell

from the late Archaic period shell deposits at Tlacuachero (level
120140 m).

Level 3.203.40

Level 4.404.60
Level (m)
Level 5.405.60

Wet

Dry

Radiocarbon
age

1.201.40

Level 6.06.20

28572464 BC

2.002.40

Level 7.07.20

3.303.40

2 3 4 5 6 7 8 9 10 11 12
18O PDB (ml1)
Figure 12. Summary of the range and mean of the 18O measurements from analysed archaeological shells from Tlacuachero (N=2
per level) and modern (N=6) specimens of Polymesoda radiata from
the Los Cerritos Lagoon. For each excavation level investigated, a
line shows the range and a dot indicates the mean of the values for
profiles of two archaeological clams. In addition, the mean and range
of the six modern shell profiles are shown.

4.404.60

34983145 BC

5.405.60

30942902 BC

6.06.20

30372788 BC

7.07.20
11 10

The range of oxygen isotopic values of all analysed

archaeological and modern shells is similar (Figure 12).
Marsh clam shells from the late Archaic period deposits exhibit mean oxygen isotopic values between
"7 and "8. The mean oxygen isotopic value of
all modern shells was "7. In general, the amplitude
of annual oxygen isotopic variation in the archaeological shells is comparable to the modern forms. Archaeological P. radiata shells collected from the uppermost
deposits (120140 m; c. 1800 ) at the late Archaic
period site of Tlacuachero deviate slightly from this
pattern (Figures 12 and 13). Dry season conditions
apparently were similar, but the lower values during
the wet season (2 lower) suggest greater rainfall than
today.
Shell margin samples from each level (N=20 per
level) are plotted in Figure 14 against the entire range
of oxygen isotopic variability exhibited by the modern
and archaeological specimens. A clear trend occurs
toward lower 18O values during the late Archaic
period. Oxygen isotopic values of terminal shell edge
samples in the lowest three levels extend across the full

18O isotope (PDB)

Figure 14. Summary of shell margin samples from each level
analysed at Tlacuachero (N=20 per level) plotted against the full
oxygen isotopic range determined from the modern and archaeological studies. -: One shell margin.

range of isotopic variability, but concentrate between

"6 and "4 (dry season). The oxygen isotopic
values in shells from the upper portion of the sequence
tend to cluster between "7 and "11 (wet season).
This culminates with a clear focus between "11 and
"10 (wet season) in the stratum just prior to late
Archaic period site abandonment.

Discussion
The analyses of modern P. radiata indicate that
changes in the oxygen isotopic composition of shell
carbonate accurately reflect fluctuations in estuarine salinity caused by seasonal changes in rainfall.
Comparison of the oxygen isotopic (18OSMOW)

700 D. J. Kennett and B. Voorhies

composition of estuarine water with 18O of shell

margin samples indicates that the carbonate precipitated by the mollusc is in isotopic equilibrium with the
water. Changes in the oxygen isotopic composition of
the water and shell margin carbonate are clearly related to fluctuations in estuarine salinity. Estuarine
salinity fluctuations are controlled by the intensity of
rainfall on the coastal plain and bordering Sierra
Madre de Chiapas mountains, although there is a lag
of approximately 3 months before the changes in
rainfall affect estuarine salinity. Presumably the lag
between rainfall and estuarine salinity occurs because
the influx of freshwater continues following the cessation of the monsoonal rains and lags at the beginning
of the following wet period until the water table rises
on the coastal plain.
Prehistoric shells from all stratigraphic levels at
Tlacuachero (CAP-7) exhibit oxygen isotopic profiles
similar to modern specimens. This suggests that the
patterns of rainfall were much the same throughout the
late Archaic period (30001800 ) and also similar to
those of today. Like most hunters and gatherers, the
people in this region during the late Archaic period
must have been affected by changes in resource availability, both the terrestrial and littoral environments.
Large fluctuations in the intensity and timing of rainfall
would have caused significant changes in the distribution of resources and we expect that humans would
have responded to them. The bedded deposits at
Tlacuachero and other littoral shell mounds of this age
appear to be uniform and are thought to be indicative of
a relatively stable subsistence strategy during the late
Archaic period (Michaels & Voorhies, 1989). Stable
patterns of middle Holocene rainfall in this region must
have contributed to the stability of human subsistence
and settlement strategies.
Significant changes in rainfall did not occur at the
end of the late Archaic period before the cessation of
shell accumulation. However, it is clear that populations living on the Pacific coastal plain during the
succeeding Formative period did change their subsistence strategies (Blake et al., 1992a), combining maize
agriculture with hunting and gathering wild resources.
In the Mazatn region people were hunting game
animals, fishing the freshwater swamp and gathering
wild plant foods. These activities were complemented
by maize horticulture. Simultaneously, higher levels of
sociopolitical complexity began to develop (Clark,
1991, 1994) on the coastal plain. The results of the
study reported here suggest to us that this socioeconomic transformation was not influenced by dramatic
environmental change at the end of the late Archaic
period (Kennett & Voorhies, 1995).
Our findings suggest that patterns of late Archaic
period shellfish harvesting were more complex than
previously suspected. A dynamic change in the use of
shellfish occurred throughout this 1000 year time interval. Early inhabitants of the Pacific coastal plain
collected shellfish and other estuarine resources at all

times during the year, with a focus during dry season

months. Although later populations in the region continued to collect marsh clams throughout the year,
there was a distinct shift towards wet season collection,
apparent at the 440460 cm level and above at
Tlacuachero (Figure 14). Just prior to the termination
of late Archaic period deposition at Tlacuachero, as
judged from the surviving record,* this trend culminated with exclusive collection of marsh clams during
wet season months.
A shift in the seasonal collection of these molluscs
suggests that the use of this location evolved throughout the late Archaic period. Deith (1985) has aptly
pointed out that the season of huntergatherer site
occupation does not always correspond with seasonal
patterns of shellfish consumption. However, based
on the available archaeological evidence from
Tlacuachero, primarily the overwhelming presence of
marsh clam shells and the near absence of structures
and intensive habitation debris, it would appear that
the season of molluscan harvesting is a good measure
of when this location was occupied.
These new data allow us to reconsider the
subsistence-settlement model proposed by Michaels &
Voorhies (1989). The seasonal patterns of clam use in
the early stages of the late Archaic period could have
been created by mobile foragers occupying and reoccupying these locations at different times during
their seasonal rounds. However, it seems to us
more likely that these sites were created by collectors,
making frequent logistical forays to the estuary. By
the end of the late Archaic period it is clear that
these locations were being used logistically by collectors as special locations for extracting estuarine
resources, as proposed by Michaels & Voorhies (1989).
This more dynamic and changing model of subsistence and settlement warrants further testing and will
require investigating other site types farther inland
dating to the early and late stages of the late Archaic
period.
Intensive use of estuarine locations in dry season
months during the early part of the late Archaic period
conforms to the expected pattern based on modern
resource availability, marked by highest productivity in
the lagoons during these months. The question is, then,
why did later late Archaic period populations apparently shift away from using these locations during the
most productive time of the year? There are several
possible explanations for this observed phenomenon,
two of which we shall examine here.
One possible explanation is that the emphasis on wet
season clam collection at the end of the late Archaic
period marsh clams represents a different scheduling
pattern of the use of littoral resources compared
*The uppermost bedded shell deposits at Tlacuachero have been
removed by later people, who may have been mining them to make
lime. For this reason, the surviving record of archaeological deposits
does not document the final stages of the shell collecting adaptation.

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 701

with earlier times. Perhaps clams were being collected

mainly during the wet season months whereas other
estuarine resources were being given priority during the
dry season. For instance, shrimp exploitation could
have intensified during the dry season to such a degree
that at that time other resources (including clams) in
the estuary were ignored. If this were the case, human
populations would have visited or lived in the estuarine
environment throughout the annual cycle, but the
focus of their subsistence activities would have changed
seasonally during the year. This scenario resembles the
procurement economy of modern inhabitants of these
wetlands: people in the village of Las Palmas focus
their procurement strategies on shrimp when they are
abundant in the lagoons and on fin fish during other
times of the year. However, other lines of archaeological evidence do not support the idea that this pattern
prevailed for the Chantuto people. For example, new
site types associated with exclusive shrimp exploitation
do not appear in the littoral zone at the end of the late
Archaic period. Moreover, there is no evidence for
heightened occupation of the known shell mound
locations, such as increased frequency of domestic
architecture, as might be expected with increased sedentism at the shell mounds. Finally, Voorhies et al.
(1991) have suggested that the presence of small fish
vertebrae throughout the deposits at Tlacuachero may
be interpreted as indirect indicators of shrimp procurement. They argue that net fishing does not allow a
fisherman to target specific prey: small fish and small
crustacea are generally present in the same net catch.
Following this reasoning, if shrimp exploitation intensified at the end of the late Archaic period, a higher
percentage of small fish vertebrae would be expected in
the upper deposits at Tlacuachero and other late
Archaic period sites, compared with the lower deposits.
This is not the case.
An alternative explanation for this dramatic shift to
wet season utilization of molluscs is that a scheduling
conflict developed with other resources elsewhere in the
region. We target the initiation of farming, perhaps
with maize (Zea mays) as one crop, as worthy of
further consideration in this respect. Maize was established as a cultigen in the highlands of Mexico by 3500
years ago (Long et al., 1989; Fritz, 1994; but see
MacNeish, 1967, 1991; Flannery & Marcus, 1983;
Flannery, 1986), a date that coincides closely with
the scheduling change observed in the record at
Tlacuachero (Figure 14; c. level 440460 m). Apparently, maize was present even earlier in the Pacific
watershed of Central Panama, where microscopic ecofacts are identified in sediment core levels dating to
approximately 7000 years ago (Piperno, 1985). On the
eastern seaboard, Pohl & Pope (pers. comm.) have
discovered indicators of maize as early as 5000 years
ago. Thus, maize might have been introduced into or
developed in the Soconusco at any time during the late
Archaic period. In view of its role as a staple in the diet
of later Mesoamerican peoples, its introduction and

increasing importance could have set in motion other

subsistence changes.
The pattern of subsistence change that was manifested by the people of the Soconusco is complex and
has not yet been fully identified by prehistorians (Blake
et al., 1992a). Macrobotanical remains of cultigens,
including maize, now have been identified as early
as the Barra phase (15501400 ) deposits in the
Soconusco (Feddema, 1993). Thus, it is indisputable
that by the Barra phase, immediately after the
Chantuto phase in the area, cultigens formed a component of the subsistence of the people of the
Soconusco.
However, it is very difficult to determine the relative
importance of these cultigens at any point in time,
given the gaps in the presently known ecofactual
record (Feddema, 1993). The best data come from
studies of the bone chemistry of the people themselves
(Blake et al., 1992b; but see Ambrose & Norr, 1992).
Analysis of the osteological remains of two Chantuto
people* gave a 13C and 15N ratios closely similar to
each other and to ratios of populations eating maize as
a staple, as well as to later peoples in the Acapetahua
region whose dependency on maize can be reasonably
assumed (Blake et al., 1992b). This shows that the shell
mound builders were heavily dependent on C4 plant
foods, although which plants were being used and their
wild versus domesticated status are unknown at
present. It is noteworthy, however, that the Early
Formative people of the Mazatn region, who we
know were growing maize because its remains are
present in the archaeological record that they produced, have 13C and 15N values that indicate a much
lower utilization of C4 plants, compared with the
earlier Chantuto people of the Acapetahua region.
Thus, we have meager but consistent evidence that
the people who were occupying Tlacuachero at the
time of the clay floor construction, were heavily dependent upon plant foods. Whether these were wild plants,
cultigens or, as is more likely, a combination of the two
has yet to be determined. It is likely that maize was
initially adopted by these huntergatherers in the
Soconusco as another resource in their broad-based
subsistence economy. As subsequent genetic changes
increased the productivity of maize, it would have
become a more viable alternative to the well established huntingfishingcollecting strategy practised by
the people of this region. This is the rationale for
speculating that scheduling conflicts may have arisen
during the last part of the late Archaic period.
The potential for farming varies in the study area,
especially across the coastal plain from the sea to the
mountains. The wetlands in the Acapetahua area are
not especially well suited for maize growing because
arable land is scarce, the soils are saline, rainfall is
seasonal, and the lower reaches of the rivers are subject
*These two individuals are the only ones that have been recovered
from the Chantuto phase deposits. Both sets of remains came from
the clay floor level at Tlacuachero.

702 D. J. Kennett and B. Voorhies

to periodic flooding during the rainy season. However,

small corn patches are located today on some river
levees and old barrier islands, so it is clear that maize
can be produced in a limited way within the estuary.
Modern farmers report that the productivity of maize
and other domesticates increases as one moves from
the coast to the piedmont area, located 2040 km
inland (Voorhies, pers. comm.; Clark, 1994). The traditional farming pattern in recent times has been to
plant two or three crops annually (Clark, 1994). The
first crop is planted in May, just as the monsoonal
rains begin and harvested in August near the end of the
rainy season. The second crop is planted in September
at the end of the rainy season and harvested in
December. A third crop was planted whenever possible
at the height of the dry season and harvested in April.
In the Mazatn area the chahuites, abandoned stream
channels that are lower and moister than the surrounding terrain, are used for this purpose, but in the
Acapetahua region where chahuites are lacking, the
lower piedmont is used. In general, maize is grown in
this region virtually year-round during optimal years
using dry farming techniques.
We hypothesize that by the late stages of the late
Archaic period the horticultural activities of the
Chantuto people were such that people were becoming
increasingly tethered to their plots, which we predict
were situated on the coastal plain, between the coastal
wetlands and the piedmont. Perhaps maize was first
grown only during the dry season when wild plant
foods would have been most scarce. Or, if maize was
grown throughout the year, as it is today, it may have
been especially important for people to have protected
their fields from herbivores during the dry season when
wild plant foods would be relatively scarce and
competition among animals particularly high.
The proposed seasonal shifts in subsistence and
settlement are somewhat comparable to ethnohistorically recorded strategies practised by aboriginal groups
on the northeast coast of North America. With the
introduction of maize around 900 there was adaptive shift from hunting and gathering wild resources
to an economy that combined farming with the exploitation of estuarine and marine resources. This led to a
progressively more sedentary way of life and the
growth of settlements on the coastal plain, away from
the coast, where farming was most productive. The
inland settlements of the Beothuk of New Foundland,
located near farming plots, were almost completely
vacated during the summer months when people were
drawn to the coast to exploit coastal shellfish beds
(Snow, 1978, 1980). Only a few very young and old
people would stay behind at the inland settlements to
protect the crops during the summer. Similarly, the
Maliseets of Northern Maine and New Brunswick
would make several trips back and forth between
garden plots and prime fishing sites during the summer
months (Erickson, 1978). In both of these cases, maize
crops were planted at the beginning of the summer

and left essentially unattended until the harvest in the

fall.
Our evidence suggests that the prehistoric people
living on the Pacific coastal plain of Mexico at the end
of the late Archaic period had a subsistence strategy
that combined horticulture and the procurement of
estuarine resources. Our working hypothesis is that the
demands of farming became increasingly great during
the dry season, thus causing the rescheduling of procurement activities in the littoral zone during the wet
season.
In summary, reconstruction of the seasonal use of
molluscs through the late Archaic period using oxygen
isotopic analysis has defined an interesting change in
resource use that would have gone unrecognized using
standard archaeological procedures. There is a clear
shift from year round procurement of clams with a
focus during dry season months at the beginning of the
Tlacuachero record of the late Archaic period, to their
use exclusively during wet season months just before
the cessation of intensive use of these locations. This
shift is best understood when scheduling conflicts associated with the development of maize agriculture in
the region are considered.

Conclusions
(1) The results presented in this study indicate that
oxygen isotopic analysis of shell carbonate from estuarine molluscs can be used to reconstruct ancient
water regimes along the Pacific coast of southwestern
Mexico. Predictable changes in oxygen isotopic composition of the water in the Acapetahua Estuary occur
in accordance with changes in salinity. Lower 18O
values correlate with the influx of fresh water into the
estuary during the wet season, although a lag of about
3 months exists between the onset of monsoonal rains
and changes in estuarine salinity.
(2) Oxygen isotopic fluctuations in the estuary are
recorded and preserved in the shells of the marsh clam
Polymesoda radiata. Oxygen isotopic profiles in shells
are used to reconstruct ancient rainfall regimes and
the establishment of patterns of prehistoric mollusc
harvesting. The lag between the onset of monsoonal
rains and changes in estuarine salinity restrict these
determinations to gross seasonal, rather than monthly,
estimates.
(3) Analysis of shells from archaeological deposits at
the site of Tlacuachero indicate that patterns of rainfall
remained virtually constant during late Archaic period
(23000 to 1800 ) with a slight increase in rainfall
c. 1800 .
(4) These data indicate that there was no significant
increase or decrease in rainfall at the end of the late
Archaic period that might account for changes in use
of estuarine resources.
(5) Based on this study and other lines of archaeological evidence it is clear that early in the late Archaic

Oxygen Isotope Analysis of Shells to Detect Seasonal Use of Wetlands 703

period (230001800 ) the site of Tlacuachero was

visited by huntergatherers throughout the year with a
focus during dry season months. Although it is likely
that the huntergatherers at this time were making
logistical forays to these locations from base camps
farther inland, it is also possible that smaller, more
mobile groups were occupying and reoccupying these
sites.
(6) Throughout the late Archaic period there was a
trend towards wet season use of Tlacuachero. By the
end of this period (21800 ) the site was used exclusively during wet season months by logistically
organized collectors.
(7) The changing use of sites in the littoral zone of the
Acapetahua Estuary is attributed to scheduling conflicts that developed as maize agriculture intensified
elsewhere on the coastal plain.

Acknowledgements
Field work that resulted in the collections of clam
shells used in this study was supported by the National
Science Foundation and the National Geographic
Society, under the auspices of the Instituto Nacional de
Antropologa e Historia, Mexico. Michael Blake and
Ronald Lowe recorded some of the environmental data
from the Comisin de Agua, and Martn de los Santos
M. collected the clam and water samples that were a
crucial component of this study.
We are grateful to James Kennett for the use of his
laboratory and helpful comments during all stages of
this research. We also thank Michael DeNiro for
making the 18O determinations on water samples.
Mark Aldenderfer, John Clark, Christina Conlee,
Michael Glassow, Michael Jochim, and Phillip Walker
provided useful comments during the preparation of
this paper and Meredith Kennett assisted with many of
the graphics. The laboratory research was funded by
grants from the Humanities/Social Science Research
Grant at the University of California, Santa Barbara
and Sigma Xi.

References
Ambrose, S. H. & Norr, L. (1992). On stable isotopic data and
prehistoric subsistence in the Soconusco region. Current Anthropology 33, 401404.
Anikouchine, N. (1990). An analysis of vertebrate faunal remains
from Chiapas, Mexico: the diet of archaic and early classic people.
Masters Thesis, Department of Anthropology, University of
California, Santa Barbara, CA, U.S.A.
Bailey, G. N., Deith, M. R. & Shackleton, N. J. (1983). Oxygen
isotope analysis and seasonality determinations: limits and potential of a new technique. American Antiquity 48, 390398.
Bamforth, D. (1986). Technological efficiency and tool curation.
American Antiquity 211, 3850.
Bernstein, D. J. (1990). Prehistoric seasonality studies in coastal
southern New England. American Anthropologist 92, 96115.
Bettinger, R. L. (1991). Hunter-Gatherers: Archaeological and Evolutionary Theory. New York: Plenum Press.

Binford, L. R. (1980). Willow smoke and dogs tails: huntergatherer

settlement systems and archaeological site formation. American
Antiquity 45, 420.
Binford, L. R. (1983). In Pursuit of the Past: Decoding the Archaeological Record. New York: Thames and Hudson.
Blake, M., Chisholm, B. S., Clark, J. E. & Mudar, K. (1992a).
Non-Agricultural Staples and Agricultural Supplements: Early
Formative Subsistence in the Soconusco Region, Mexico. In (A. B.
Gebauer & T. D. Price, Eds.) Transitions to Agriculture in Prehistory. Maidson, WI: Monographs in World Archaeology No., 4.
Prehistory Press, pp. 133151.
Blake, M., Chisholm, B. S., Clark, J. E., Voorhies, B. & Love, M. W.
(1992b). Prehistoric Subsistence in the Soconusco Region. Current
Anthropology 33, 8394.
Blake, M., Clark, J. E., Voorhies, B., Michaels, G., Love, M., Pye,
M. E., Demarest, A. A. & Arroyo, B. (1995). A radiocarbon
chronology for the Late Archaic and Formative periods along the
Pacific Coast of southeastern Mesoamerica. Ancient Mesoamerica
6, 161183.
Clark, J. E. (1991). The Beginnings of Mesoamerica: Apologia for
the Soconusco Early Formative. In (W. Fowler, Jr., Ed.) The
Formation of Complex Society in Southeastern Mesoamerica, Boca
Raton, LA: CRC Press, pp. 1327.
Clark, J. E. (1994). The development of early formative ranked
societies in the Soconusco, Chiapas, Mesico. Ph.D. Thesis, Department of Anthropology, University of Michigan, Ann Arbor, MI,
U.S.A.
Deith, M. R. (1983). Seasonality of shell collecting determined by
oxygen isotope analysis of marine shells from Asturian sites in
Cantabria. In (C. Grigson & J. Clutton-Brock, Eds.) Animals and
Archaeology: 2 Shell Middens, Fishes, and Birds. Oxford: BAR
International Series 183, pp. 6776.
Deith, M. R. (1985). Seasonality from shells: an evaluation of two
techniques for seasonal dating of marine molluscs. In (N. R. J.
Fieller, D. D. Gilbertson & N. G. A. Ralph, Eds.) Palaeobiological
Investigations Research Design, Methods and Data Analysis,
Oxford: BAR International Series 266, pp. 119130.
Deith, M. R. (1988). Shell seasonality: an appraisal of the oxygen
isotope technique, recent developments. In (R. E. Webb, Ed.)
Environmental Analysis in Old and New World Archaeology
Oxford: BAR International Series 416 pp. 3749.
Drucker, P. (1948). Preliminary notes on an archaeological survey of
the Chiapas Coast. Middle American Research Records 1, 151169.
Epstein, S., Buchsbaum, R., Lowenstam, H. & Urey, H. (1951).
Carbonatewater isotopic temperature scale. Bulletin of the
Geological Society of America 62, 417426.
Epstein, S., Buchsbaum, R., Lowenstam, H. & Urey, H. (1953).
Revised carbonatewater isotopic temperture scale. Bulletin of the
Geological Society of America 64, 13151326.
Erickson, V. O. (1978). Maliseet-Passamaquoddy. In (B. G. Trigger,
Ed.) Handbook of North American Indians. Vol 15. Washington,
DC: Smithsonian Institution, pp. 123136.
Fairbanks, R. G. (1989). A 17 000-year glacio-eustatic sea level
record: influence of glacial melting rates on the younger Dryas
event and deep-ocean circulation. Nature 342, 637641.
Feddema, V. L. (1993). Early formative subsistence and agriculture
in southeastern Mesoamerica. Masters thesis. Department of
Anthropology and Sociology, University of British Columbia,
Vancouver, Canada.
Flannery, K. V. (Ed.) (1986). Guil Naquitz: Archaic Foraging and
Early Agriculture in Oaxaca, Mexico. New York: Academic Press.
Flannery, K. V. & Marcus, J. (Eds.) (1983). The Cloud People:
Divergent Evolution of Zapotec and Mixtec Civilizations. New
York: Academic Press.
Fritz, G. J. (1994). Are the first American farmers getting younger?
Current Anthropology 35, 305308.
Glassow, M. A., Kennett, D. J., Kennett, J. P. & Wilcoxon, L. R.
(1994). Confirmation of Middle Holocene ocean cooling inferred
from stable isotopic analysis of prehistoric shells from Santa Cruz
Island, California. In (W. L. Halvorson & G. J. Maender, Eds.)

704 D. J. Kennett and B. Voorhies

The 4th Californian Islands Symposium: Update on the Status of
Resources. Santa Barbara, CA: Santa Barbara Museum of Natural
History.
Hard, R. J. & Merrill, W. L. (1992). Mobile agriculturalists and the
emergence of sedentism: perspectives from northern Mexico.
American Anthropologist 94, 601620.
Herz, N. (1990). Stable isotope geochemistry applied to archaeology.
In (N. P. Lasca & J. Donahue, Eds.) Archaeological Geology of
North America, Vol. 3. Boulder, CO: Geology Society of America,
pp. 585595.
Hudson, J. L., Walker, P. L. & Voorhies, B. (1989). Changing
patterns of faunal exploitation. In (B. Voorhies, Ed.) Ancient
Trade and Tribute: Economies of the Soconusco Region of Mesoamerica. Salt Lake City: University of Utah Press, pp. 135155.
Jones, J. G. (1988). Archaeobotanical analyses of sediments from
the Tlacuachero site, Chiapas Mexico. Manuscript on file, Palynology Laboratory, Department of Anthropology, Texas A & M
University, College Station, TX, U.S.A.
Jones, J. G. & Mora, G. (1993). Analysis of pollen and phytolith
remains from the Lesher site, Chiapas, Mexico. Manuscript on file,
Department of Anthropology, University of Colorado, Boulder,
CO, U.S.A.
Keith, M. L., Anderson, G..M. & Eichler, R. (1993). Carbon and
oxygen isotopic composition of mollusk shells from marine and
fresh-water environments. Geochimica et Cosmochimica Acta 28,
17871816.
Kennett, D. J. & Voorhies, B. (1995). Middle Holocene periodicities
in rainfall inferred from oxygen and carbon isotopic fluctuations
in prehistoric tropical estuarine mollusc shells. Archaeometry 37,
157170.
Killingley, J. S. (1990). Seasonality of mollusk collecting at Hubbs
Midden 1959: VI: 28A. Pacific Coast Archaeological Society
Quaterly 16, 1923.
Killingley, J. S. (1981). Seasonality of mollusk collecting determined
from 18O profiles of midden shells. American Antiquity 48, 152
158.
Killingley, J. S. & Berger, W. H. (1979). Stable isotopes in a mollusk
shell: detection of upwelling events. Science 205, 186188.
Lieberman, D. E. (1993). The rise and fall of seasonal mobility
among huntergatherers: the case of the southern Levant. Current
Anthropology 34, 599632.
Long, A., Benz, B. F., Donahue, D. J., Jull, A. J. T. & Toolin, L. J.
(1989). First direct AMS dates of early maize from Tehuacn,
Mexico. Radiocarbon 31, 10351040.
Lorenzo, J. L. (1955). Los Concheros de la Costa Chiapas. Anales del
Instituto Nacional de Antropologa e Historia 7, 4150.
MacNeish, R. S. (1967). A summary of subsistence. In (D. Byers,
Ed.) The Prehistory of the Tehuacn Valley Environment and
Subsistence. Vol. 1. Austin, TX: University of Texas Press,
pp. 290309.
MacNeish, R. S. (1991). The Origins of Agriculture and Settled Life.
Norman, OK: University of Oklahoma Press.
Michaels, G. H. & Voorhies, B. (1989). Late Archaic period coastal
collectors in southern Mesoamerica: the Chantuto people revisited. Proceedings of the Circum-Pacific Prehistory Conference,
Developments of HuntingFishingGathering Maritime Societies on
the Pacific. Seattle, Washington, August 26 1989.

Monks, G. G. (1981). Seasonality Studies. In (M.,B. Schiffer, Ed.)

Advances in Archaeological Method and Theory. Vol. 4. New York:
Academic Press, pp. 117240.
Navarrete, C. (n.d.). Resumen de las exploraciones de reconocimiento
arqueolgico de la Costa de Chiapas (Regin del Soconusco), en la
Temporada de 1969. Manuscript on file, New World Archaeological Foundation, San Cristobal de las Casas, Mexico.
Piperno, D. (1985). Phytolithic analysis of geological sediments from
Panama. Antiquity 59, 1319.
Shackleton, N. J. (1969). Marine mollusca in archaeology. In Science
in Archaeology, (D. Brothwell & E., S. Higgs, Eds.) New York:
Thames and Hudson, pp. 407414.
Shackleton, N. J. (1973). Oxygen isotope analysis as a means of
determining season of occupation of prehistoric midden sites.
Archaeometry 15, 133141.
Snow, D. R. (1978). Late prehistory of the East Coast. In (B. G.
Trigger, Ed.) Northeast, Handbook of North American Indians.
Vol. 15. Washington, DC: Smithsonian Institution, pp. 5869.
Snow, D. R. (1980). The Archaeology of New England. New York:
Academic Press.
Severyn, N. J. (1993). Taxonomic revision and phylogeny of the genus
Polymesoda (Bivalvia: Corbiculidae). Ph.D. Thesis, MEES Program, University of Maryland, Eastern Shore Campus, Princess
Anne, MA, U.S.A.
Torrence, R. (1983). Time budgeting and huntergatherer technology. In (G. Bailey, Ed.) HunterGatherer Economy in Prehistory:
A European Perspective. Cambridge: Cambridge University Press,
pp. 1122.
Urey, H. C. (1947). The thermodynamic properties of isotopic
substances. Journal of the Chemical Society 1947, 562581.
Viv Escoto, J. A. (1964). Weather and climate of Mexico and
Central America. In (Robert Wauchope, Ed.) Handbook of Middle
American Indians, Vol. 1. Austin, TX: University of Texas Press,
pp. 216264.
Voorhies, B. (1976). The Chantuto People: An Archaic Period Society
of the Chiapas Littoral, Mexico. Papers of the New World Archaeological Foundation. Vol. 41. Provo: Brigham Young University.
Voorhies, B. (1991). The transformation from foraging to farming in
the lowlands of Mesoamerica. In (S. L. Fedick, Ed.) The Managed
Mosaic: Ancient Maya Agriculture and Resource Use. Salt Lake
City, UT: University of Utah Press.
Voorhies, B. (Ed.) (1989). Ancient Trade and Tribute: Economies of
the Soconusco Region of Mesoamerica. Salt Lake City, UT:
University of Utah Press.
Voorhies, B. & Kennett, D. J. (1994). Buried Sites on the Soconusco
Coastal Plain. Journal of Field Archaeology 22, 6579.
Voorhies, B., Michaels, G. H. & Riser, G. M. (1991). Ancient
shrimp fishery. National Geographic Research and Exploration 7,
2035.
Waselkov, G. A. (1987). Shellfish gathering and shell midden archaeology. In (M. B. Schiffer, Ed.) Advances in Archaeological Method
and Theory, Vol. 10. New York: Academic Press, pp. 93210.
Wefer, G. & Berger, W. H. (1991). Isotope paleontology: growth and
composition of extant calcareous species. Marine Geology 100,
207248.